New·GLM-5.2 / Kimi K3

One gateway for production AI traffic

Unified model access with routing controls, real-time usage, secure keys and predictable billing.

Explore features
API endpoint
https://api.caiaiu.com
China mainland endpoint
https://cn.caiaiu.com

One endpoint. Every frontier model.

The infrastructure layer between your product and the model providers.

openai·anthropic·gemini

Smart routing controls

Route traffic across OpenAI, Anthropic, Gemini and more through one endpoint, with provider changes handled behind the gateway.

usage

Usage you can see

Live request, token and duration metrics by provider, so spend never surprises you.

prepaid·metered·no min

Pay as you go

Top up a balance or subscribe. Transparent metered usage, payment limits and spend visibility without monthly minimums.

cai_sk_····a91f

Scoped API keys

Mint, rotate and revoke keys per project. Every key carries its own limits and audit trail.

rpm·spend cap·24h window

Production guardrails

Credentials stay server-side, with rate limits, scoped keys and usage windows to reduce abuse and runaway spend.

base_url=api.caiaiu.com/v1

Drop-in compatible

Speaks the OpenAI API shape. Point your existing SDK at our base URL and you are done.

Every frontier model, one endpoint

Approximate daily request allowances under the Pro plan. Live usage is governed by your balance and active plan limits.

These estimates reflect the Pro plan’s daily quota, not all plans.

Daily usage estimate

Token-cost estimate
200Kimi K3
700GLM-5.2
800Kimi K2.7 Code
1,500DeepSeek V4 Pro
2,000DeepSeek V4 Flash
4,500MiniMax M33x usage

Reference request counts converted from token-cost assumptions for a Pro subscription. Actual usage varies with input/output length, cache use, model pricing, balance, and active plan limits.

See it in motion

One key, every model, live usage and drop-in SDK access — the whole gateway in 33 seconds.

Frequently asked